It is well known that the gonadotropin-releasing hormone (GnRH) pulse generator is highly sensitive to environmental insults, particularly weight loss and low body fat and we propose that the modulation of GnRH is mediated by a leptin-metabolic axis. Anorexia nervosa in young women is associated with severe weight loss, amenorrhea and a leptin deficient state, which reverses with weight gain. This problem appears to be due to a hypothalamic dysfunction affecting GnRH pulsatility. Those who do not experience a reversal generally demonstrate disordered eating. The calorically depressed state of anorexia nervosa is also associated with a decrease in resting metabolic rate (RMR). Recent work on obesity suggests that the RMR may remain depressed in individuals who maintain a body weight that is lower than their 'normal' weight. This continues to be a problem with recovered anorectics. We hypothesize that anorexia nervosa and associated disordered eating are closely tied to a depressed RMR and a metabolic cascade reflecting inadequate metabolic resources which influences the secretion of leptin and the expression of GnRH. An increase in the RMR will be associated with a return of normal leptin levels and GnRH function, resulting eventually in normal cyclicity. We propose that those who continue to experience amenorrhea will continue to have low RMR, low leptin levels and manifest disordered eating. A detailed study of young women with anorexia nervosa and their patterns of disordered eating serve as a perfect model to investigate the link between nutrition, leptin, RMR and the neuroendocrine and ovarian axes. This study will measure RMR using a unique state of the art chamber, changes in reproductive hormones and leptin levels, GnRH pulsatility (LH and FSH pulse studies), responses to GnRH and leptin and cortisol pulses in women receiving treatment for anorexia nervosa. Thirty-six patients and fourteen controls will be followed with adrenal hormone profiles, hormonal markers of nutritional intake, body composition and eating disorder profiles. Data will be obtained before and after a refeeding protocol which brings patients to 90% of ideal body weight, and at six month intervals for a follow-up period of 1 year. Understanding the dysfunction seen in anorexia nervosa could provide a key to the long sought metabolic signal involved in GnRH pulsatility, an important central mechanism in the reproductive cycle of women, as well as open up new avenues for treatment.